Study On The Simulation Of Multispectral Infrared Images

The simulation of infrared multispectral images is one of leading techniques in modern infrared field. As a combination of infrared physics, heat transfer, computer graphics and virtual reality, the simulation of infrared multispectral images supplies a virtual platform to an infrared multispectral imaging system. This thesis is concerned with improving infrared simulation results that based on the Vega simulation platform. In traditional Vega infrared simulation methods, the temperature of an object is derived from one-dimensional heat conduction function, which results in a high uncertainty from a real physic process. The proposed approach applied a three-dimensional heat conduction function based on Vega and ANSYS infrared multispectral imagery simulation, and produced a superior infrared simulation result in terms of simulation accuracy. This thesis presented two infrared multispectral image simulation methods based on Vega platform. Both of them could generate infrared multispectral images under varied imaging conditions including different seasons, weather and time.This thesis introduced transmission properties of the infrared radiation in the atmosphere, analyzed effects resulting from a variety of atmosphere ingredients on infrared radiation transmission at different bands, and derived equations to quantify these effects respectively. Atmosphere data used in this thesis were obtained by setting imaging conditions in MAT. Several core steps in the implementation process of infrared imagery simulation were also discussed. Due to its conveniences, Vega platform was mainly utilized for infrared multispectral imagery simulation. We also briefly introduced the Vega platform, the process of Pro/Engineer object modeling as well as MultiGen Creator background modeling, discussed a model of regular temperature filed and solutions that applied ANSYS for finite-element analysis, finally, provided a framework of infrared multispectral sensor system as well as ways of implementation.This thesis presented two infrared multispectral imagery simulation methods based on Vega platform in details: One was directly transferring a visible light image into 4 sub-band infrared images ranging from 8~12μm. This method avoided a complex 3D modeling process and was superior in terms of effectiveness. Another method modeled the whole scene through Pro/Engineer object modeling and MultiGen Creator background modeling, and then imported images into Vega platform for infrared multispectral imagery simulation through OpenFlight API format. Experiments indicated that the presented method was more close to a real physic process.The final part of this thesis provided a conclusion of the whole research as well as future directions.